Comments on Dodson et al., A Technical Comparison of Evaluating Asbestos Concentration by Phase Contrast Microscopy (PCM), Scanning Electron Microscopy (SEM), and Analytical Transmission Electron Microscopy (ATEM) as Illustrated from Data Generated from a Case Report

To the Editor:

We read with interest this study by Dodson et al. (2008). Unfortunately, the analysis by these authors indicates a profound misunderstanding of the role of fiber analysis in the determination of causation in asbestos-related diseases as well as a lack of appreciation of the biases and limitations of their own technique. We take issue with a number of statements made by the authors throughout the paper, but for the sake of brevity, will confine our comments to the Summary at the end of the article.

The authors claim that evaluation of lung tissue by ATEM at 15,000x identified a significant elevated concentration of chrysotile asbestos that was not found by SEM at 1,000x. This statement is simply inaccurate. As clearly stated in our previous publication (Butnor et al., 2003), case 1 (also analyzed by Dodson et al.) contained elevated levels of both commercial amphibole fibers (amosite) and chrysotile. The exact numbers of fibers detected by the two methods are irrelevant as shown by the international interlaboratory counting trial (participated in by our laboratory, but not Dodson's) as reported by Gylseth et al. (1985). This latter study showed marked variation in actual numbers of fibers reported when the same sample was analyzed by different laboratories. What is relevant is the relationship between fiber levels of a given type in a particular case to levels in a reference unexposed population analyzed by the same laboratory (Henderson et al., 1997). In this regard, the findings by Dodson et al. (2008) are no different than the findings that we reported.

The authors state that their article describes in great detail the issues that are encountered in accurately determining asbestos concentrations in lung tissue. However, they seem to be unaware of the study by Roberson et al. (1992), which showed that ATEM is biased towards the detection of chrysotile fibers, whereas SEM is biased towards the detection of commercial amphibole fibers (e.g., amosite). Furthermore, scientifically rigorous studies have shown that the detection of chrysotile fibrils less than one micron in length is inherently inaccurate, with a 50:50 chance of different trained observers detecting fibrils in this size range when counting the same sample (Steel and Small, 1985). In spite of the limitations of both techniques, there is notable agreement between SEM and ATEM in the analysis of fiber burdens in patients with mesothelioma. The study of 54 cases of mesothelioma from various regions of the USA by Dodson et al. (2005) came to similar conclusions as our analysis of more than 400 such cases (Roggli et al., 2008), i.e. that somewhere between 80 and 90% of mesothelioma cases have a fiber burden in excess of background ranges, and amosite is the predominant fiber type identified.

Finally, the authors state that our technique would miss a significant number of short fibers, which could result in conclusions that are not factually correct with respect to causation of a given asbestos-related disease. Dodson's writings (Dodson, 2006; Dodson et al., 2003 and 2008) consistently ignore the classic inhalation studies of J.M.G. Davis and colleagues (Davis and Jones, 1988; Davis et al., 1986), which failed to demonstrate pathogenic effects for fibers less than five microns in length. These observations were confirmed in the extensive reanalysis of Davis' studies by Berman et al. (1995). An expert panel convened by the Agency for Toxic Substances and Disease Registry came to a similar conclusion in a meeting in New York in October, 2002. The great majority of investigators in asbestos-related diseases have moved on from the ‘short-fiber’ debate, but Dodson and colleagues inexplicably persist (vested interest, perhaps?) in their concerns about the pathogenicity of such fibers.

In summary, we believe that the findings in the two cases reported by Dodson et al. (2005 and 2008; the second case was not referred to in their most recent publication) support one of the main conclusions which we derived from the ten cases that we reported (Butnor et al., 2003) and the five additional cases that we have studied since that time (unpublished observations), which is that among mesothelioma patients for whom the only recognized source of asbestos exposure was to friction products, the finding of an increased level of chrysotile asbestos in tissue samples is invariably associated with increased levels of commercial amphiboles (amosite or crocidolite). Whereas products containing commercial amphiboles frequently also contain chrysotile, commercial amphiboles have not been used in friction products in the USA. Taken together, these findings cast serious doubt on any claims that increased levels of chrysotile in such a patient derived from friction products.